4.3BSD-Reno/src/sys/hpdev/fd.c
/*
* Copyright (c) 1988 University of Utah.
* Copyright (c) 1990 The Regents of the University of California.
* All rights reserved.
*
* This code is derived from software contributed to Berkeley by
* the Systems Programming Group of the University of Utah Computer
* Science Department.
*
* Redistribution is only permitted until one year after the first shipment
* of 4.4BSD by the Regents. Otherwise, redistribution and use in source and
* binary forms are permitted provided that: (1) source distributions retain
* this entire copyright notice and comment, and (2) distributions including
* binaries display the following acknowledgement: This product includes
* software developed by the University of California, Berkeley and its
* contributors'' in the documentation or other materials provided with the
* distribution and in all advertising materials mentioning features or use
* of this software. Neither the name of the University nor the names of
* its contributors may be used to endorse or promote products derived from
* this software without specific prior written permission.
* THIS SOFTWARE IS PROVIDED AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED
* WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
*
* from: Utah $Hdr: fd.c 1.3 89/12/03$
*
* @(#)fd.c 7.1 (Berkeley) 5/8/90
*/
/*
* File (vnode) disk driver.
*
* Block/character interface to a vnode. Note that this uses the
* VOP_BMAP/VOP_STRATEGY interface to the vnode instead of a simple
* VOP_RDWR. We do this to avoid distorting the local buffer cache.
*
* NOTE: There is a security issue involved with this driver.
* Once mounted all access to the contents of the "mapped" file via
* the special file is controlled by the permissions on the special
* file, the protection of the mapped file is ignored (effectively,
* by using root credentials in all transactions).
*/
#include "fd.h"
#if NFD > 0
#include "param.h"
#include "systm.h"
#include "buf.h"
#include "errno.h"
#include "dkstat.h"
#include "ioctl.h"
#include "user.h"
#include "vfs.h"
#include "vnode.h"
#include "file.h"
#include "uio.h"
#include "malloc.h"
#include "fdioctl.h"
#ifdef DEBUG
int fddebug = 0x00;
#define FDB_FOLLOW 0x01
#define FDB_INIT 0x02
#define FDB_IO 0x04
#endif
struct buf fdbuf[NFD];
struct buf fdtab[NFD];
#define b_cylin b_resid
#define fdunit(x) ((minor(x) >> 3) & 0x7) /* for consistency */
#define getfdbuf() \
((struct buf *)malloc(sizeof(struct buf), M_DEVBUF, M_WAITOK))
#define putfdbuf(bp) \
free((caddr_t)(bp), M_DEVBUF)
struct fd_softc {
int sc_flags; /* flags */
size_t sc_size; /* size of fd */
struct vnode *sc_vp; /* vnode */
struct ucred *sc_cred; /* credentials */
int sc_maxactive; /* max # of active requests */
} fd_softc[NFD];
/* sc_flags */
#define FDF_ALIVE 0x01
#define FDF_INITED 0x02
fdopen(dev, flags)
dev_t dev;
{
int unit = fdunit(dev);
#ifdef DEBUG
if (fddebug & FDB_FOLLOW)
printf("fdopen(%x, %x)\n", dev, flags);
#endif
if (unit >= NFD)
return(ENXIO);
return(0);
}
/*
* Break the request into bsize pieces and submit using VOP_BMAP/VOP_STRATEGY.
* Note that this driver can only be used for swapping over NFS on the hp
* since nfs_strategy on the vax cannot handle u-areas and page tables.
*/
fdstrategy(bp)
register struct buf *bp;
{
int unit = fdunit(bp->b_dev);
register struct fd_softc *fs = &fd_softc[unit];
register struct buf *nbp;
register int bn, bsize, resid;
register caddr_t addr;
int sz, flags;
extern int fdiodone();
#ifdef DEBUG
if (fddebug & FDB_FOLLOW)
printf("fdstrategy(%x): unit %d\n", bp, unit);
#endif
if ((fs->sc_flags & FDF_INITED) == 0) {
bp->b_error = ENXIO;
bp->b_flags |= B_ERROR;
iodone(bp);
return;
}
bn = bp->b_blkno;
sz = howmany(bp->b_bcount, DEV_BSIZE);
bp->b_resid = bp->b_bcount;
if (bn < 0 || bn + sz > fs->sc_size) {
if (bn != fs->sc_size) {
bp->b_error = EINVAL;
bp->b_flags |= B_ERROR;
}
iodone(bp);
return;
}
bn = dbtob(bn);
bsize = fs->sc_vp->v_vfsp->vfs_bsize;
addr = bp->b_un.b_addr;
flags = bp->b_flags | B_CALL;
for (resid = bp->b_resid; resid; resid -= sz) {
struct vnode *vp;
daddr_t nbn;
int off, s;
nbp = getfdbuf();
off = bn % bsize;
sz = MIN(bsize - off, resid);
(void) VOP_BMAP(fs->sc_vp, bn / bsize, &vp, &nbn);
#ifdef DEBUG
if (fddebug & FDB_IO)
printf("fdstrategy: vp %x/%x bn %x/%x dev %x\n",
fs->sc_vp, vp, bn, nbn, vp->v_rdev);
#endif
nbp->b_flags = flags;
nbp->b_bcount = sz;
nbp->b_bufsize = bp->b_bufsize;
nbp->b_error = 0;
nbp->b_dev = vp->v_rdev;
nbp->b_un.b_addr = addr;
nbp->b_blkno = nbn + btodb(off);
nbp->b_proc = bp->b_proc;
nbp->b_iodone = fdiodone;
nbp->b_vp = vp;
nbp->b_pfcent = (int) bp; /* XXX */
/*
* Just sort by block number
*/
nbp->b_cylin = nbp->b_blkno;
s = splbio();
disksort(&fdtab[unit], nbp);
if (fdtab[unit].b_active < fs->sc_maxactive) {
fdtab[unit].b_active++;
fdstart(unit);
}
splx(s);
bn += sz;
addr += sz;
}
}
/*
* Feed requests sequentially.
* We do it this way to keep from flooding NFS servers if we are connected
* to an NFS file. This places the burden on the client rather than the
* server.
*/
fdstart(unit)
{
register struct fd_softc *fs = &fd_softc[unit];
register struct buf *bp;
/*
* Dequeue now since lower level strategy routine might
* queue using same links
*/
bp = fdtab[unit].b_actf;
fdtab[unit].b_actf = bp->b_actf;
#ifdef DEBUG
if (fddebug & FDB_IO)
printf("fdstart(%d): bp %x vp %x blkno %x addr %x cnt %x\n",
unit, bp, bp->b_vp, bp->b_blkno, bp->b_un.b_addr,
bp->b_bcount);
#endif
VOP_STRATEGY(bp);
}
fdiodone(bp)
register struct buf *bp;
{
register struct buf *pbp = (struct buf *)bp->b_pfcent; /* XXX */
register int unit = fdunit(pbp->b_dev);
int s;
s = splbio();
#ifdef DEBUG
if (fddebug & FDB_IO)
printf("fdiodone(%d): bp %x vp %x blkno %x addr %x cnt %x\n",
unit, bp, bp->b_vp, bp->b_blkno, bp->b_un.b_addr,
bp->b_bcount);
#endif
if (bp->b_error) {
#ifdef DEBUG
if (fddebug & FDB_IO)
printf("fdiodone: bp %x error %d\n", bp, bp->b_error);
#endif
pbp->b_flags |= B_ERROR;
pbp->b_error = geterror(bp);
}
pbp->b_resid -= bp->b_bcount;
putfdbuf(bp);
if (pbp->b_resid == 0) {
#ifdef DEBUG
if (fddebug & FDB_IO)
printf("fdiodone: pbp %x iodone\n", pbp);
#endif
iodone(pbp);
}
if (fdtab[unit].b_actf)
fdstart(unit);
else
fdtab[unit].b_active--;
splx(s);
}
fdread(dev, uio)
dev_t dev;
struct uio *uio;
{
register int unit = fdunit(dev);
#ifdef DEBUG
if (fddebug & FDB_FOLLOW)
printf("fdread(%x, %x)\n", dev, uio);
#endif
return(physio(fdstrategy, &fdbuf[unit], dev, B_READ, minphys, uio));
}
fdwrite(dev, uio)
dev_t dev;
struct uio *uio;
{
register int unit = fdunit(dev);
#ifdef DEBUG
if (fddebug & FDB_FOLLOW)
printf("fdwrite(%x, %x)\n", dev, uio);
#endif
return(physio(fdstrategy, &fdbuf[unit], dev, B_WRITE, minphys, uio));
}
/* ARGSUSED */
fdioctl(dev, cmd, data, flag)
dev_t dev;
u_long cmd;
caddr_t data;
int flag;
{
int unit = fdunit(dev);
register struct fd_softc *fs;
struct fd_ioctl *fio;
struct vattr vattr;
struct vnode *vp;
int error;
#ifdef DEBUG
if (fddebug & FDB_FOLLOW)
printf("fdioctl(%x, %x, %x, %x): unit %d\n",
dev, cmd, data, flag, unit);
#endif
error = suser(u.u_cred, &u.u_acflag);
if (error)
return (error);
if (unit >= NFD)
return (ENXIO);
fs = &fd_softc[unit];
fio = (struct fd_ioctl *)data;
switch (cmd) {
case FDIOCSET:
if (fs->sc_flags & FDF_INITED)
return(EBUSY);
/*
* Always open for read and write.
* This is probably bogus, but it lets vn_open()
* weed out directories, sockets, etc. so we don't
* have to worry about them.
*/
error = vn_open(fio->fd_file, UIO_USERSPACE,
FREAD|FWRITE, 0, &vp);
if (error)
return(error);
error = VOP_GETATTR(vp, &vattr, u.u_cred);
if (error) {
vn_close(vp, FREAD|FWRITE);
VN_RELE(vp);
return(error);
}
fs->sc_vp = vp;
fs->sc_size = btodb(vattr.va_size); /* note truncation */
error = fdsetcred(fs);
if (error) {
vn_close(vp, FREAD|FWRITE);
VN_RELE(vp);
return(error);
}
fdthrottle(fs, vp);
fio->fd_size = dbtob(fs->sc_size);
fs->sc_flags |= FDF_INITED;
#ifdef DEBUG
if (fddebug & FDB_INIT)
printf("fdioctl: SET vp %x size %x\n",
fs->sc_vp, fs->sc_size);
#endif
break;
case FDIOCCLR:
if ((fs->sc_flags & FDF_INITED) == 0)
return(ENXIO);
fdclear(fs);
#ifdef DEBUG
if (fddebug & FDB_INIT)
printf("fdioctl: CLRed\n");
#endif
break;
default:
return(ENXIO);
}
return(0);
}
/*
* Duplicate the current processes' credentials. Since we are called only
* as the result of a SET ioctl and only root can do that, any future access
* to this "disk" is essentially as root. Note that credentials may change
* if some other uid can write directly to the mapped file (NFS).
*/
fdsetcred(fs)
register struct fd_softc *fs;
{
struct uio auio;
struct iovec aiov;
char tmpbuf[DEV_BSIZE];
fs->sc_cred = crdup(u.u_cred);
/* XXX: Horrible kludge to establish credentials for NFS */
aiov.iov_base = tmpbuf;
aiov.iov_len = MIN(DEV_BSIZE, dbtob(fs->sc_size));
auio.uio_iov = &aiov;
auio.uio_iovcnt = 1;
auio.uio_offset = 0;
auio.uio_rw = UIO_READ;
auio.uio_segflg = UIO_SYSSPACE;
auio.uio_resid = aiov.iov_len;
return(VOP_READ(fs->sc_vp, &auio, 0, fs->sc_cred));
}
/*
* Set maxactive based on FS type
*/
fdthrottle(fs, vp)
register struct fd_softc *fs;
struct vnode *vp;
{
extern struct vnodeops ufs_vnodeops, nfs_vnodeops;
if (vp->v_op == &nfs_vnodeops)
fs->sc_maxactive = 2;
else
fs->sc_maxactive = 8;
if (fs->sc_maxactive < 1)
fs->sc_maxactive = 1;
}
fdshutdown()
{
register struct fd_softc *fs;
for (fs = &fd_softc[0]; fs < &fd_softc[NFD]; fs++)
if (fs->sc_flags & FDF_INITED)
fdclear(fs);
}
fdclear(fs)
register struct fd_softc *fs;
{
register struct vnode *vp = fs->sc_vp;
#ifdef DEBUG
if (fddebug & FDB_FOLLOW)
printf("fdclear(%x): vp %x\n", vp);
#endif
fs->sc_flags &= ~FDF_INITED;
if (vp == (struct vnode *)0)
panic("fdioctl: null vp");
#if 0
/* XXX - this doesn't work right now */
(void) VOP_FSYNC(vp, fs->sc_cred);
#endif
vn_close(vp, FREAD|FWRITE);
VN_RELE(vp);
crfree(fs->sc_cred);
fs->sc_vp = (struct vnode *)0;
fs->sc_cred = (struct ucred *)0;
fs->sc_size = 0;
}
fdsize(dev)
dev_t dev;
{
int unit = fdunit(dev);
register struct fd_softc *fs = &fd_softc[unit];
if (unit >= NFD || (fs->sc_flags & FDF_INITED) == 0)
return(-1);
return(fs->sc_size);
}
fddump(dev)
{
return(ENXIO);
}
#endif